Experimental Comparisons of Deep Neural Network and Machine Learning Lung Cancer Detection Algorithms for CT Images

Worldwide, the number of people who pass away as a result of lung cancer is on the rise. Due to advancements in AI technology testing for cancer at an earlier stage can save various lives. Lung cancer detection methods that entail many steps and necessitate human empirical parameter adjustments are often employed. This research paper presents a comparative study between machine and deep learning models. First lung slice and nodule segmentation using ML and DL is essential for cancer detection. Deep learning techniques have improved healthcare image analysis. There are a few machine learning approaches like random forest, decision tree, logistic regression, and SVM along with deep learning approaches like CNN, ResNet50,101,50V2, VGG16,19, and transfer learning: DenseNet201 and InceptionV3 over chest CT dataset. Experimental results on chest CT images have demonstrated that logistic regression (ML) and transfer learning (DL) can provide superior performance to the state-of-the-art methods regarding accuracy and loss decrement.